在mysql中选择动态列

时间:2012-08-10 10:18:38

标签: mysql sql pivot

是否可以遍历这样的表:

mysql> select * from `stackoverflow`.`Results`;
+--------------+---------+-------------+--------+
| ID           | TYPE    | CRITERIA_ID | RESULT |
+--------------+---------+-------------+--------+
|            1 | car     | env         | 1      |
|            2 | car     | gas         |        |
|            3 | car     | age         |        |
|            4 | bike    | env         | 1      |
|            5 | bike    | gas         |        |
|            6 | bike    | age         | 1      |
|            7 | bus     | env         | 1      |
|            8 | bus     | gas         | 1      |
|            9 | bus     | age         | 1      |
+--------------+---------+-------------+--------+
9 rows in set (0.00 sec)

进入这个:

+------+-----+-----+-----+
| TYPE | env | gas | age |
+------+-----+-----+-----+
| car  | 1   |     |     |
| bike | 1   |     | 1   |
| bus  | 1   | 1   | 1   |
+------+-----+-----+-----+

目的是选择所有CRITERIA_ID并将其用作列。 作为行我喜欢使用所有TYPE s。

  • 所有标准:SELECT distinct(CRITERIA_ID) FROM stackoverflow.Results;
  • 所有类型SELECT distinct(TYPE) FROM stackoverflow.Results;

但是如何将它们组合成视图或smth。像这样?

如果你喜欢玩数据。这是一个生成表格的脚本:

CREATE SCHEMA `stackoverflow`;
CREATE TABLE `stackoverflow`.`Results` (
  `ID` bigint(20) NOT NULL AUTO_INCREMENT,
  `TYPE` varchar(50) NOT NULL,
  `CRITERIA_ID` varchar(5) NOT NULL,
  `RESULT` bit(1) NOT NULL,
  PRIMARY KEY (`ID`)
) 
ENGINE=InnoDB;

INSERT INTO `stackoverflow`.`Results`
(
 `ID`,
 `TYPE`,
 `CRITERIA_ID`,
 `RESULT`
)
VALUES
( 1, "car", env, true ),
( 2, "car", gas, false ),
( 3, "car", age, false ),
( 4, "bike", env, true ),
( 5, "bike", gas, false ),
( 6, "bike", age, true ),
( 7, "bus", env, true ),
( 8, "bus", gas, true ),
( 9, "bus", age, true );

1 个答案:

答案 0 :(得分:10)

不幸的是,MySQL没有PIVOT功能,这基本上就是你要做的。因此,您需要使用带有CASE语句的聚合函数:

SELECT type,
  sum(case when criteria_id = 'env' then result end) env,
  sum(case when criteria_id = 'gas' then result end) gas,
  sum(case when criteria_id = 'age' then result end) age
FROM results
group by type

请参阅SQL Fiddle with Demo

现在,如果您想要动态执行此操作,这意味着您事先不知道要转置的列,那么您应该查看以下文章:

Dynamic pivot tables (transform rows to columns)

您的代码如下所示:

SET @sql = NULL;
SELECT
  GROUP_CONCAT(DISTINCT
    CONCAT(
      'SUM(IF(CRITERIA_ID = ''',
      CRITERIA_ID,
      ''', RESULT, NULL)) AS ',
      CRITERIA_ID
    )
  ) INTO @sql
FROM
  Results;
SET @sql = CONCAT('SELECT type, ', @sql, ' FROM Results GROUP BY type');

PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

请参阅SQL Fiddle with Demo